Describing Repeat Abortions in Eurasia: A Comparison of Women in Kazakhstan, Uzbekistan and Kyrgyzstan

Jamie L. Trevitt, Johns Hopkins University
Nan M. Astone, Johns Hopkins University

We explore the demographic determinants of repeat abortion in Kazakhstan, Uzbekistan and Kyrgyzstan using data from Demographic Health Surveys. The outcome variable of interest is the number of abortions a woman had up to the date of the interview. We employ the novel method of generalized ordered logit regression with multiple covariates, which is similar to an ordinal logit model but does not rely on the assumption of proportional odds. We find that while there are significant factors that differ across these countries, being Muslim not only has proportional odds in each country model but is also consistently associated with lower odds of repeat abortion. Years of sexual exposure does not have proportional odds for each country model, but does consistently have a positive effect on the number of abortions a woman has had. Our results suggest that unmet need for contraception remains a serious health problem in Eurasia.

  See paper

Presented in Session 48: Abortion Prevalence, Measurement and Programs